Hi folks! My name is Dima and I am a PhD student at UCSC. I do machine learning research. My advisor is Manfred Warmuth. Bulk of my work has been on online algorithms, specifically on generalizations that learn matrix parameters. Related to this is my work on generalizing probability theory to density matrices, which are most commonly used in quantum physics. Another direction included work on compression schemes. Size of a compression scheme is a parameter characterizing the complexity of a learning task.

My thesis is almost done and I'll be joining Google to help them with organizing the world's information and all of that good stuff! :D

You can contact me by e-mailing dima "doggy sign" soe dt ucsc dt edu

Here is a list of my publications so far:

1. Kuzmin D. and Warmuth M.K. Online Kernel PCA with Entropic Matrix Updates, ICML 2007. [pdf]
2. Warmuth M.K. and Kuzmin D. Randomized PCA Algorithms with Regret Bounds that are Logarithmic in the Dimension, NIPS 2006. [pdf]
3. Warmuth M.K. and Kuzmin D. A Bayesian Probability Calculus for Density Matrices, UAI 2006 . [pdf]
4. Warmuth M.K. and Kuzmin D. Online Variance Minimization, COLT 2006 . [pdf]
5. Kuzmin D. and Warmuth M.K. Unlabeled Compression Schemes for Maximum Classes, Journal of Machine Learning Research, accepted for publication, 2007. [pdf]
6. Kuzmin D. and Warmuth M.K. Unlabeled Compression Schemes for Maximum Classes, COLT 2005. [pdf]
7. Kuzmin D. and Warmuth M.K. Optimum Follow the Leader Algorithm, COLT 2005 open problem. [pdf]